From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:0.9.3) Gecko/20010801 BuildID: 2001080110 The search engine at the above URL generates URLs which contain ampersands. It uses & in the URL for each occurance. There is a sherlock plugin for this engine at search.cpg.com/sherlock/cpga.src . When generated URLs containing & occur in the sidebar search panel tab, the & is not converted to & (and so the link is broken), whereas in the Mozilla Search panel they are converted (and the link is good). Reproducible: Always Steps to Reproduce: 1. Install the plugin search.cpg.com/sherlock/cpga.src 2. Set the sidebar search tab preference to advanced 3. Perform a search for 'java' using the newly added engine and at least one other engine Actual Results: The results of this search should include several 'category' results, e.g., 'JavaClass'. If you select this link in the search tab, notice the loaded URL contains '&' strings. If you then select the same link in the search results pane, the URL contains '&' chars instead (but sometimes does not even load the page). Expected Results: The URLs should be consistent, hopefully with '&' converted to '&'. Additional Info: This occurs in a Solaris/SPARC build as well.

->sgehani/097/p3

Moving to mozilla0.9.8.

Moving to milestone after mozilla0.9.9 (mozilla1.0 for now).

Search triage team: -> Future

I think this is a duplicate of 152038.

*** Bug 152038 has been marked as a duplicate of this bug. ***

Moving information from duped bug 152038 to this one (target milestone as set by sgehani, summary which is much clearer and CC to mat)

*** Bug 216783 has been marked as a duplicate of this bug. ***

Decodes HTML entities in href attribute. For regular links Mozilla does this anyway. This patch removes a line where CR/LF were removed. The called function also does this.

Comment on attachment 137502 [details] [diff] [review] patch: decode HTML entities Requesting r= from timeless and sr= from dbaron. And... that patch is ... "inspired" by timeless, not really my work...

Comment on attachment 137502 [details] [diff] [review] patch: decode HTML entities Well, it's short, it's correct (urls in href tags are supposed to be parsed for entities, the old code didn't), and it's tested. This is a low risk and localized patch. The payoff is that we could ship an ask jeeves sherlock plugin.

Comment on attachment 137502 [details] [diff] [review] patch: decode HTML entities a=asa (on behalf of drivers) for checkin to Mozilla 1.6

Checking in xpfe/components/search/src/nsInternetSearchService.cpp; /cvsroot/mozilla/xpfe/components/search/src/nsInternetSearchService.cpp,v <-- nsInternetSearchService.cpp new revision: 1.199; previous revision: 1.198 done